30th ANNIVERSARY ISSUE

Evaluating handwashing technique Elaine Larson

PhD

Robert Wood Johnson Clinical Nurse Scholar, University of Pennsylvania School of Nursing, Pennsylvania, USA

Ed Lusk Director, Biostatistics, Cancer Centre, Hospital of the University of Pennsylvania, Pennsylvania, USA

Accepted for publication 18 March 1985

Correspondence: Elaine Larson, Johns Hopkins University, School of Nursing, 600 N. Wolfe Street, Baltimore, MD 21205, USA.

Journal of Advanced Nursing 10, 547–552 L A R S O N E . & L U S K E . ( 2 0 0 6 ) Journal of Advanced Nursing 53(1), 46–53 Evaluating handwashing technique Though standards for handwashing have been defined, little effort has been made to assess the quality of handwashing in clinical settings. This paper describes tests of reliability and validity of tools to evaluate two aspects of handwashing — appropriateness and technique. Based on these tests, methods to evaluate handwashing are recommended. LARSON E. & LUSK E. (1985)

Introduction Though many studies in the past few decades have sought to correlate handwashing frequency and agent used with the numbers and types of bacteria on the skin of the hands (Larson 1984, Aly & Maibach 1976, Casewell & Phillips 1977, Ojajarvi 1980), very little has been done to assess the quality or technique of handwashing as practised in clinical settings. Those few published studies that have reported the rating of handwashing technique have used observation tools for which validity and reliability data are unavailable (Fox et al. 1974, Taylor 1978). The purpose of our studies reported here was to develop standardized tools for the evaluation of handwashing technique in the clinical setting.

Review of the literature Guidelines published by Centres for Disease Control (CDC) in 1981 stated that handwashing should always occur: 1 before performing invasive procedures; 2 before and after contact with wounds; 3 before contact with susceptible patients; and 4 after contact with a source likely to be contaminated. The guidelines stated that handwashing was not required after brief, routine direct patient contacts such as shaking hands or measuring blood pressure (Simmons 1981). This is a modification of earlier recommendations by CDC and the 46

American Hospital Association (AHA) that hands be washed before and after contact with each patient (Centres for Disease Control 1970, American Hospital Association 1974). There seems to be an attempt in these new guidelines to provide staff with recommendations that are more directed at high risk patient contacts. The guidelines also require a higher level of understanding on the part of the practitioner, who must now decide whether the preceding or succeeding patient contact is of a magnitude or type that warrants handwashing. It is quite possible that the guidelines are more congruent with clinical practice than the more simplistic directive that the hands should be washed before and after every patient contact. In the 1970s, Fulkerson at CDC developed a scale ranking patient and environmental contacts for their relative potential as contaminants of the hands (Fox et al. 1974). This scale was subsequently used in several observational studies to determine whether or not handwashing should have occurred (Fox et al. 1974, Taylor 1978). In these studies some categories were classified as ‘clean’ and some as ‘dirty’. This clean-dirty division seems compatible with the new CDC guidelines. Using this scale, Taylor found in 129 observations of handwashing that nurses did not distinguish between clean and dirty situations. She hypothesized that this was due to an affective problem: ‘[Nurses] appear to believe that if their hands are not physically soiled...then no spread of infection can occur’ (Taylor 1978). It was not possible to

Ó 1985 Blackwell Science Ltd Ó 2006 The Authors. Journal compilation Ó 2006 Blackwell Publishing Ltd

30th anniversary issue Table 1 Ranking activities. 1. Sterile or autoclaved materials 2. Thoroughly cleaned or washed materials 3. Materials not necessarily cleaned but free from patient contact (e.g., papers, nursing station) 4. Objects contacted by patients either infrequently or not expected to be contaminated (e.g., patient furniture) 5. Objects intimately associated with patients, but not known to be contaminated (e.g., patient gowns, linens, dishes, bedside rails) 6. Minimal, limited contact with patient (e.g., shaking hands, taking pulse) 7. Objects in contact with patient secretions 8. Patient contact in which secretions or mouth, nose, rectum, etc. are touched 9. Materials contaminated by patient urine 10. Direct contact with patient urine 11. Materials contaminated with faeces 12. Direct contact with faeces 13. Materials contaminated with secretions or excretions from infected sites 14. Direct contact with secretions or excretions from infected sites 15. Direct contact with infected patient sites (e.g., wounds, tracheotomy) ‘Clean’ activities: 1–7 ‘Dirty’ activities: 8–15

determine whether the nurses’ failure at handwashing was a result of an inadequate knowledge and understanding of levels of contamination or of an attitude about the relative cleanliness of their hands. Unfortunately, tests of reproducibility and validity of this tool were not reported (see copy of adapted ranking tool in Table 1). In addition to the tool ranking patient care activities from cleanest to dirtiest, a second tool developed by Feldman (1969) has been used to evaluate the quality or technique of handwashing. Though it contained all of the essential components of handwashing defined by the Association for Practitioners in Infection Control (Soule 1983), it had not been tested for reproducibility between observers or for sensitivity. Clinical evaluation of handwashing practices, then, is not a simple task. There are five components which together comprise the quality of a handwashing event: frequency, agent used, appropriateness (i.e. whether the hands were washed when they should have been), duration and technique. In this study we are reporting the development and testing of tools to assess the appropriateness and technique of handwashing.

Methods Studies were conducted at the Hospital of the University of Pennsylvania.

Appropriateness A questionnaire was developed to evaluate the extent to which the clean-dirty rankings on Fulkerson’s scale correlated with rankings assigned by health care personnel. The 15

clean-to-dirty items on the scale were listed in random order on the questionnaire. Hospital personnel were asked to rank the items from cleanest to dirtiest. To determine whether there was agreement regarding the point of contamination at which handwashing should occur, they were also asked to indicate each item which should definitely be followed by handwashing. A convenience sample of physicians, nurses and nursing auxilliary personnel was used. Tests for differences in proportions and Chi square tests of homogeneity (Armitage 1971) were used to test hypotheses.

Handwashing technique In our initial clinical trials of Feldman’s scale for observing handwashing technique, we found a number of problems. Wording was subject to individual interpretation and scores obtained by simultaneous observers were inconsistent and not reproducible. Therefore, we undertook a series of tests. First the tool was modified for clarity and ease of scoring. Ambiguous words were changed and definitions were made to maximize objectivity. For example, vigorous friction was defined as including visible movement of both arms up to the elbow and an audible sound of rubbing (unless the flow of water was too loud to allow one to hear the sound). Minimal friction was defined as the movement of hands, but not arms. No friction was defined as an absence of rubbing of hands together after soap was applied. We attempted to make categories mutually exclusive and of equal importance and weight. The tool was then sent to 10 experts in infection control throughout the country (five metropolitan areas). Experts were chosen because of their activity in the Association of Practitioners in Infection Control and/or their publications in the field of infection

Ó 1985 Blackwell Science Ltd, Journal of Advanced Nursing, 10, 547–552 Ó 2006 The Authors. Journal compilation Ó 2006 Blackwell Publishing Ltd

47

30th anniversary issue

control. These experts were asked to weight the relative importance of each of the components of the tool to confirm that each component was indeed independent and of equal importance (in which case a particular score, say 10, could always be interpreted as demonstrating an equivalent level of handwashing quality). When the tool had been reviewed by experts it was again tested for inter-rater reliability in the hospital by several independent observers: three nurses from different clinical areas and a laboratorian. Because there was very little range in the quality of handwashing observed during that testing period (it was all about ‘average’), a second set of tests, again with three observers, was conducted to establish whether the tool would be sensitive in differentiating between poor, average, and good handwashing, and whether differences within each of the seven categories would be consistently recognized. In these tests, individuals who washed their hands were instructed to make specific errors. They randomly selected a piece of paper on which instructions for handwashing were written. For example, ‘wash your hands normally, but hold your hands up during rinsing so that water drips up your arms’ or ‘wash your hands normally, but don’t use friction. Just put the soap on and rinse it off’. The observers, of course, were ignorant of the instructions to the handwasher.

Results

Table 2 Number of responses regarding the need for handwashing (HW) among physicians (n ¼ 32), nurses (n ¼ 90) and nursing assistants and laboratorians (n ¼ 8).

HW No HW Totals Proportions

MD

RN

Other

Totals

370 83 480 0Æ771

999 366 1365 0Æ732

87 18 105 0Æ829

1483 467 1950 0Æ761

Chi square test, 2 d.f. ¼ 14Æ42, P < 0Æ001 One nurse did not respond to this section of the questionnaire

direct contact with known infected secretions or excretions (item 13), contact with materials contaminated by patient urine (item 9), and contact with materials soiled with faeces (item 11). The mean and median ranks assigned by physicians and nurses agreed ±2 points with the original ranking for all 15 items. The mean and median ranks assigned by nursing assistants and laboratorians on all items but one were within 2 points of the original rank order. That one discrepancy was contact with materials contaminated with secretions or excretions from infected sites (original ranking was 13; mean and median for these groups were 10Æ2 and 10Æ7 respectively). Physicians and nursing assistants indicated the need for handwashing significantly more often than did nurses (77Æ1% for physicians, 82Æ9% for nursing assistants and laboratorians, 73Æ2% for nurses, P < 0Æ001, see Table 2).

Appropriateness A total of 255 questionnaires were distributed to hospital staff through the mail or in person; 131 (51Æ4%) were returned and useable. Response rates by position were: physicians, 31Æ4% (32/102); registered nurses, 64Æ1% (91/ 142); and nursing assistants and laboratorians, 72Æ7% (8/11). Personnel responding came from several clinical areas: adult intensive care, 29 (22Æ1%); obstetrics and newborn nursery, 24 (18Æ3%); medicine, 24 (18Æ3%); nursing faculty, 19 (14Æ5%); surgery, 12 (9Æ2%); infection control, 11 (8Æ4%); and oncology, 10 (7Æ6%). Two (1Æ5%) individuals did not specify their clinical area. The entire group indicated significantly more often that handwashing was needed following the eight ‘dirty’ exposures (99Æ2%) than following the seven ‘clean’ contacts (46Æ6%); test of proportions P < 0Æ0001. All individuals (100%) responded that handwashing was indicated after direct contact with infected patient sites (item 15 on scale) and after direct contact with faecal soilage (item 12). All but one respondent indicated that handwashing should follow contact with secretions or excretions known to be contaminated (item 14), contact with materials which had been in 48

Technique The handwashing observation tool in its final form is presented in Table 3. The majority of experts felt that the seven categories were not of equal importance and that some items should be assigned additional weight; this is reflected in the final tool. We found that the tool was very sensitive in identifying all errors in technique which were intentionally made by subjects. In our final tests, 20/21 (95Æ2%) of observations performed simultaneously by three individuals yielded total scores within two points of each other.

Discussion The response rates we obtained were typical of patterns seen with voluntary, mailed questionnaires. Since physicians have been shown in several studies to report and to practise handwashing significantly less than nurses (Larson 1981, Albert & Condie 1981), it is surprising that in this study they identified the need for handwashing significantly more often than did nurses. One cannot rule out the possibility that this finding might be explained by differences in response rates in

Ó 1985 Blackwell Science Ltd, Journal of Advanced Nursing, 10, 547–552 Ó 2006 The Authors. Journal compilation Ó 2006 Blackwell Publishing Ltd

30th anniversary issue Table 3 Handwashing observation. Total duration* Score 1 Soap Visible bubbles No visible bubbles No soap 2 Splashing No splashing on clothes or floor Minimal splashing on clothes or floor Vigorous splashing on clothes or floor 3 Friction (rubbing) Vigorous friction (visible arm movement and/or audible running sounds) Minimal friction (hands only moving) 4 Surfaces covered Dorsal, palmar, interdigital areas covered Two of above surfaces covered One surface only covered 5 Hand position Hands held down so water drains from fingers Hands held parallel with arms 6 Rinse All surfaces rinsed Only parts of hands rinsed No rinsing 7 Drying Dried hands thoroughly, turned off faucet with paper towel Dried hands, turned off faucet with hands (unless knee, foot or elbow controlled sink) Did not dry hands Total score:

Weight

Maximum total

2 1 0

x

2

4

2 1 0

x

1

2

x

2

4

x

2

4

x

1

2

2 1 0

x

1

2

2 1 0

x

1

2

2 1 2 1 0 2 1

20

*Duration ¼ time from initial placing of soap on the hands to rinsing off the soap.

the two groups (the physicians who responded to the questionnaire might have felt more strongly about handwashing). In addition, the numbers of nursing assistants and laboratorians was too small to make inferences about that group. Nevertheless, the overwhelming ability of this group of health care personnel to differentiate between clean and dirty activities and to accurately rank their potential ‘cleanliness’ or ‘dirtiness’ makes it clear that these individuals can accurately identify a need for handwashing. This study did not address the issue of changes in handwashing behaviour caused by the presence of the observer. There is some evidence that behaviours like handwashing which are habits that are practised frequently are changed very little under observation (McLane et al. 1983). There are two ways to deal with this problem; observe handwashing clandestinely, or make observations over an extended period of time so that subjects will become accustomed to the presence of the observer. Others have observed handwashing without the knowledge of subjects (Fox et al. 1974, Taylor 1978, Albert & Condie 1981). We

are currently conducting a study during which handwashing observations are made over a number of months. A second issue of concern is the method by which frequency of handwashing is determined. Most studies to date have used subject report of frequency. In a current study we are correlating subject report of handwashing frequency with observed frequency to evaluate the reliability of people’s reported estimates. Reports of these studies will appear in subsequent publications.

Summary This study reports the development of two tools for evaluating handwashing technique. The above results demonstrate the accuracy and reliability of the ranking scale as a method for determining the need for, or appropriateness of, handwashing in clinical settings. In view of the high degree of reproducibility attained with the modifications described in this study, it is suggested that others studying handwashing test these tools in their clinical settings.

Ó 1985 Blackwell Science Ltd, Journal of Advanced Nursing, 10, 547–552 Ó 2006 The Authors. Journal compilation Ó 2006 Blackwell Publishing Ltd

49

30th anniversary issue

References Albert R.K. & Condie F. (1981) Handwashing patterns in medical intensive care units. New England Journal of Medicine 304, 1465– 1466. Aly R. & Maibach H.I. (1976) Effect of antimicrobial soap containing chlorhexidine an the microbial flora of skin. Applied Environmental Microbiology 31, 931–935. American Hospital Association (1974) Infection Control in the Hospital 3rd edn. American Hospital Association, Chicago. Armitage P. (1971) Statistical Methods in Medical Research. Blackwell Scientific, Oxford. Casewell M. & Phillips I. (1977) Hands as route of transmission for Klebsiella species. British Medical Journal 2, 1315–1317. Centres for Disease Control. (1970) Isolation Techniques for use in Hospitals. Government Printing Office, Washington DC. Feldman H. (1969) Learning transfer from programmed instruction to clinical performance. Nursing Research 18, 51–54. Fox M.K., Langner S.B. & Wells R.W. (1974) How good are handwashing practices? American Journal of Nursing 74, 1676–1678.

Larson E.L. (1981) Persistent carriage of gram-negative bacteria on hands. American Journal of Infection Control 9, 112–119. Larson E.L. (1984) Effects of handwashing frequency, agent used, and clinical unit on bacterial colonization of the hands. American Journal of Infection Control 12, 76–82. McLane C., Chenelly S., Sylvestrak M.C. & Kirchhoff K.T. (1983) A nursing practice problem: failure to observe aseptic technique. American Journal of Infection Control 11, 178–182. Ojajarvi J. (1980) Effectiveness of handwashing and disinfection methods in removing transient bacteria after patient nursing. Journal of Hygiene 85, 193–203. Simmons B.P. (1981) Antiseptics, Handwashing, and Handwashing Facilities. Centres for Disease Control, Atlanta. Soule B.M. (1983) The APIC Curriculum for Infection Control Practice, vol I. Kendall/Hunt, Dubuque. Taylor L.J. (1978) An evaluation of handwashing techniques – 2. Nursing Times 74, 108–110.

30th anniversary commentary on Larson E. & Lusk E. (1985) Evaluating handwashing technique. Journal of Advanced Nursing 10, 547–552 Elaine Larson

PhD RN FAAN CIC

Professor of Pharmaceutical and Therapeutic Research, Associate Dean for Research, Columbia University School of Nursing, Professor of Epidemiology, Joseph Mailman School of Public Health, Columbia University, Editor, American Journal of Infection Control, 630 W. 168th St, New York, NY 10032, USA. E-mail: [email protected]

A never-ending story Since the late 1970s my collaborators and I have been studying the impact of hand hygiene on healthcare-associated infections. This programme of research has continued into the present time; a few examples of recent publications arising from our work are included in the reference list (Larson et al. 2005a, 2005b, 2005c). It was during my postdoctoral work at the University of Pennsylvania in the mid-1980s, however, that it became clear that there were no validated tools to examine either the appropriateness of, or the technique used, for handwashing. The 1985 Journal of Advanced Nursing paper, written with a statistician, Ed Lusk, represented one of my first forays into development and psychometric testing of several simple instruments. The instruments which are described in that paper, however, are now outdated because the Centres for Disease 50

Control and Prevention (CDC) in the USA has published a new guideline for hand hygiene, now recommending use of waterless alcohol hand sanitizers, rather than traditional handwashing techniques for patient care (Boyce & Pittet 2002). But to me there are still two valuable lessons that can be taken from the 1985 Journal of Advanced Nursing paper. First, regardless of the starting point for a programme of research, eventually it is necessary to branch out and acquire or seek expertise in a variety of fields. For me, this has meant drawing on microbiology, dermatology, psychometrics and biostatistics, decision theory and sociology, to name just a few fields. Secondly, even with what appears to be the simplest of research topics – what could be simpler than handwashing?! – it turns out that there is a never-ending stream of new knowledge to describe and apply. Sadly, but happily too, the ‘definitive study’ will never be done. This is what makes research so dynamic and interesting.

Ó 2006 The Author. Journal compilation Ó 2006 Blackwell Publishing Ltd